Some known electronic devices use speech recognition. Electronic devices utilising voice commands and speech-to-text functionality can enable hands-free use and better accessibility. For example, a mobile phone with hands-free functionality enables a user to make a phone call or dictate a text message while driving a car.
One known way for a user to start a speech recognition function in an electronic device is to press a button or select the feature from a menu of the electronic device's user interface. This means that a user typically has to activate the speech recognition functionality with a hand or even both hands. However, a user may not be able to touch the electronic device to activate a speech recognition function. For example, a user may be washing dishes or riding a bicycle and does not want to touch or cannot touch the electronic device.
One proposal is to configure an electronic device so that a speech recognition function is on all the time. However, this may not be feasible because a processor may require large amounts of processing capacity which can consume too much power and drain a power cell of the electronic device too quickly. For example, a processor running a speech recognition function can draw a current in the order of a hundred milliamps from the power cell when calculating fast Fourier transform and processing audio to determine whether the audio contains spoken words.
In another known arrangement a signal is captured with an in-ear microphone, which detects a specific user humming sequence or a voice command. The ear piece attenuates surrounding noises from entering the ear canal making the user's voice to be easier to separate from other sounds. However, the detection of the voice command may similarly drain the power cell of an electronic device as quickly as permanently on speech recognition.
In one known arrangement a speech recognition function is activated when a user claps or whistles. Typically, the electronic device activates the speech recognition function when the user's clapping or whistling is over a certain volume. The arrangement is problematic and can lead to unnecessary activation of the speech recognition function, especially in noisy environments.
Embodiments of the application aim to address one or several of the above issues.